Brazilian court blocks Meta from using news content to train AI

The São Paulo Court Ruling

The 44th Civil Court of São Paulo issued a preliminary injunction on November 3, 2026, targeting Meta's plans for news content. The decision was prompted by a lawsuit filed by three major Brazilian media associations: the national newspaper association ANJ, the broadcasters’ association Abert, and the magazine publishers’ association Aner.

The court order specifically prevents Meta from collecting or using the content of these member outlets without explicit authorization. This ban is intended to stop the company from using published journalism to develop and train its artificial intelligence systems. The associations argue that the move addresses clear indications of contractual abuse and copyright infringement.

To ensure compliance, the court established a significant financial penalty. Meta faces a fine of R$1 million per day if it fails to follow the order, with a total cap of R$100 million. The injunction remains in effect until the case reaches a final judgment, effectively stalling Meta's planned commercial terms.

Meta's Disputed Terms

The legal battle centers on Meta's updated commercial terms for businesses using Facebook and Instagram. These terms were originally intended to let Meta use automated tools to crawl websites linked to business products. While Meta removed the specific word "train" from an early version of the text, publishers argue that the remaining language still covers generative AI development.

Under the proposed terms, publishers would be considered to have accepted the crawling clause simply by continuing to use Meta's business products after receiving notice. The media groups argue that there is no practical way to refuse these terms without completely leaving the platform. This creates a situation where publishers feel forced to give up their content rights to maintain their digital presence.

Meta has suggested that content owners could use robots.txt files to signal their preferences regarding crawling. However, trade reports indicate that no version of the new terms actually mentions the use of robots.txt. The scheduled start date for these terms was moved from October 30 to November 20, though Meta noted the date could slip if consultations with industry groups continue.

A Growing Global Trend

Brazil is not the only country taking action against Meta's AI data practices. In July 2024, Brazil's data protection authority suspended a different policy change that would have allowed Meta to train AI on personal user data. That specific case resulted in Meta suspending its generative AI features within the country to remain compliant with local laws.

The current 2026 case is distinct because it focuses on published journalism rather than personal data. It represents a shift toward media organizations as the primary plaintiffs seeking to protect their commercial assets. The associations describe this ruling as the first known decision of its kind for a coalition of media groups seeking to block AI training.

Similar actions are occurring in Europe and other regions. A Norwegian trade body filed for an injunction on behalf of ten media groups in late October, and the Oslo District Court issued a temporary block on the same terms. Additionally, the European Publishers Council has asked the European Commission to intervene, while the UK competition authority is currently in urgent talks with Meta.

What Happens Next

The immediate next step involves watching whether Meta chooses to appeal the São Paulo order or issue a public response. Because the decision is preliminary, the legal battle is far from over. The associations view this ruling as a significant first step in a much larger legal fight over the ownership of digital content in the age of AI.

Meta has already signaled that it may delay the implementation of its new terms if consultations with industry groups run past the November 20 deadline. The company's communications chief stated that they do not plan to make policy changes without full consultation. This could provide a window for negotiations between the tech giant and the media associations.

The outcome of the Oslo case in Norway and potential actions by the European Commission will also be critical. These international developments will determine if Brazil's ruling is an isolated incident or part of a coordinated global effort to regulate AI training data. For now, the Brazilian media groups have successfully secured a temporary shield for their content.